Backbone double-mutant cycle analysis quantifies hydrogen-bond energies in proteins

Protein structure is stabilized by a variety of noncovalent interactions, but many remain incompletely understood. For example, several potentially ubiquitous interactions involving the protein backbone have been identified, but challenges in determining reliable experimental energies have prevented their integration into structural models. To address this challenge, we adapted a popular protein engineering approach, double-mutant cycle analysis, for the quantification of backbone interactions. By combining this analytical paradigm with chemical peptide synthesis, we selectively probe backbone interactions while avoiding many confounding factors that have complicated previous efforts. We first validate this approach by quantifying the energy of canonical, cross-strand hydrogen bonds in model β-sheet proteins and find excellent agreement with previous results. We then extend this approach to quantify weak, intra-strand hydrogen bonds that have recently been implicated in protein folding and misfolding. Our results provide the first experimental quantification of these interactions, corroborating previous computational predictions that individual intra-strand hydrogen bonds contribute approx. 0.2 kcal/mol each, which is significantly given the frequency of these interactions. More broadly, our results illustrate a useful approach to probing backbone interactions in proteins, which can be readily applied to a variety of other systems.